Large Scale Control System Theory and Applications

大型控制系统及其应用：控制系统和自动化领域研究的是如何改变系统的行为，从而使系统发生一些“令人满意的”事情。例如，在一家化工厂中，有100‘S的测量阀和100’S的控制阀可以变化，要求该工厂能够生产一些所需的输出产品，例如聚合物。然而，在这种情况下，由于工厂行为的复杂性，以及由于始终存在无法测量的干扰的事实，不可能有一个人甚至一组人能够正确地调节工厂的阀门(输入)以实现所需的产品。为了实现控制工厂的这一目标，需要使用“控制器”，它通常是计算机上的一组指令。“简单的工厂”通常有一个“集中控制器”，其中系统的所有信息都被输入到计算机中，但是，对于像化工厂这样的复杂系统，使用这样一个集中控制器和一个“分散控制器”是不现实的，“分散控制器”由许多子控制器组成，每个子控制器都是相互隔离的。然而，问题是“分散控制”理论仍然不完善，分散控制设计还处于起步阶段。研究建议：本研究项目是申请者对所谓的“大系统控制理论”研究的继续。特别是关于分散控制，特别是分散控制器的设计。新理论提出：根据正在发展的新理论，重点是：(I)澄清仍未很好理解的分散固定模式(DFM)的性质；(Ii)尝试获得分散控制的基本限制；(Iii)尝试一种新的对象分散控制器设计方法，该方法具有不需要数学模型的特性；认为这种设计方法很重要，因为在工业系统中很难进行建模。(4)研究了各种复杂的分散控制设计方法。提出了新的应用问题：结合新的理论，很有兴趣确定该理论在应用于现实世界的问题时有多好，因此一些应用问题也将被研究，这些问题本身就很有趣。(I)将设计一种用于新一代硬盘驱动器系统的控制器，据信将比现有的磁盘驱动器快一个数量级(美国加利福尼亚州圣何塞的西部数据公司、高级伺服机械技术公司对这种新控制器设计感兴趣；注意：似乎没有加拿大公司可以与之互动。)(2)将为包括多个分布式能源(DER)单元的微电网系统的自主模式运行实施大型通用电力管理系统和分散控制器。这种类型的问题在微电网系统中很重要，将比以前的研究要求高一个数量级。(Iii)将开展开发一种适用于脊髓损伤患者的控制器的问题。(4)研究第三代大型柔性空间结构(LFSS)的分散控制，特别是存在解的充要条件，并对控制器进行表征。这一理论将在MSAT II LFSS中得到验证。

LARGE SCALE CONTROL SYSTEMS AND APPLICATIONSBACKGROUND:The field of Control Systems and Automation deals with the problem of modifying the behaviour of a system, so that something "desirable" occurs. For example in a chemical factory, which has literally 100's of measurements and 100's of control valves to vary, it is desired that the plant should produce some desired output product , e.g. a polymer. In this case howeverit is impossible to have a single human or even a group of humans to be able to correctly adjust the valves (inputs) of the plantto achieve the desired product due to the complexity of the plant's behaviour, and due to the fact that disturbances, which areunmeasurable, are always present. To achieve this objective of controlling a plant, a "controller" is used, which typically is a set of instructions on a computer. "Simple plants" generally have a "centralized controller" in which all information of the system is fed into the computer but, for complex systems such as a chemical plant, it is not realistic to have such a centralized controller and a " decentralized controller",which consists of many sub-controllers, each of which are all isolated from each other, is used. The problem however, is that the theory of "decentralized control" is still incomplete, and decentralized control design is still in its infancy.RESEARCH STUDY PROPOSED:The research project proposed is a continuation of previous studies made by the applicant on so called "Large scale control systemstheory". In particular the focus is about decentralized control, and in particular the design of decentralized controllers.NEW THEORY PROPOSED:In terms of the new theory being developed, a focus on: (i) Clarifying the properties of decentralized fixed modes (DFM), which are still not well understood, will be made(ii) An attempt on obtaining the fundamental limitations of decentralized control will be made (iii) A new decentralized controller design method for plants will be attempted which has the property that it requires no mathematical model; it is believed that such a design approach is important, because model building is difficult to carry out in industrial systems.(iv) A study of various complex decentralized control design methods will be made. NEW APPLICATION PROBLEMS PROPOSED:In conjunction with new theory, it is of great interest to determine how well the theory actually "works" when applied to realworld problems, and so some application problems will also be studied, which are interesting in their own right.(i) A controller for new generation hard disc drive systems will be designed which is believed will be an order of magnitude faster than existing disc drives (Western Digital Corp., Advanced Servo-Mechanical Technologies, San Jose, CA, USA are interested in this new controller design; note:there appears to be no Canadian company to interact with.) (ii) A large generalized electric power management system and decentralized controller for the autonomous mode operation of a microgrid system that includes multiple distributed energy resource (DER) units will be carried out. This type of problem is important in microgrid systems and will be an order of magnitude more demanding than previous studies.(iii) The problem of developing a controller which works "well" for the case of patients with spinal cord injury will be carried out. (iv) A study of the decentralized control of third generation large flexible space structures (LFSS) will be made; in particular necessary and sufficient conditions for a a solution to exist and a controller characterization will be made. The theory will be tested out on applying it to the MSAT II LFSS.